Ultra-wideband Microstrip Filter Modeling And Design Based On Genetic Algorithm

With the rapid development of modern communication technology, wireless communication technology is moving towards high-speed, multiple-band, and large-capacity. Microstrip bandpass filter is playing an important role in modern communication radio frequency circuits and it is always widely used due to the advantages of simple structure, easy maintenance, easy to integration, and low cost, thus having wide development potential and application prospect. Therefore, high performance, miniaturization, and multi-band filter design and theoretical analysis have gained much attention in microwave and millimeter wave circuit research area, especially the research on modeling and design method of microstrip bandpass filter which is becoming hot topics currently.In this thesis, a variation of genetic algorithm (VGA) is proposed and innovatively applied to intelligent modeling and design of microstrip bandpass filter. The variation of genetic algorithm overcomes the two possible drawbacks of conventional genetic algorithm (GA), i.e. slow rate of convergence and local-best solution. According to the predetermined filter performance index, in the case of no predetermined filter structure, the variation of genetic algorithm is applied to simultaneously search for appropriate circuit topology and the corresponding electrical parameters of the desired filter. This method is also not rely on prior knowledge to obtain the optimal circuit topology and its corresponding electrical parameters. To demonstrate the validity of the proposed theory, the novel compact triple-band bandpass filter (BPF) and ultra-wideband (UWB) bandpass filter (BPF) are presented and designed. On the basis, the UWB BPF with multiple notched bands design and theoretical analysis have carried on more in-depth research, and a series of new UWB BPF with dual/triple notched bands are proposed and designed. The designed theory has advantages of enhancing the automation, easy to implement, greatly shorten the cycle of the modeling and design of microstrip BPF. What is more, the designed theory improves the efficiency and accuracy of the BPF design and solves many design challenges. In conclusion, the proposed variation of genetic algorithm has very strong theory significance and engineering application value.